vito@trwspf.UUCP (Herb Barad) (01/09/87)
I remember reading that some people were having problems with
patch #3 that was posted for the 2.11 news. Is there any official
suggestions as to going ahead with patch #3. I think somewhere
people were mentioning that patch #3 could clear out the
active file. Is this true?
--
Herb Barad - TRW Data Systems Lab
ARPA: barad@brand.usc.edu or vito%trwspf.uucp@brand.usc.edu
USENET: ...!{brand|trwrb}!trwspf!vitorick@seismo.CSS.GOV (Rick Adams) (01/12/87)
If you are running 4.[23] BSd or a system with lockf() you can install it with no problems. if you are using links for the file locking, a bit of carelessness on my part in integrating several different fixes introduced a potential problem (potential since all compilers so far have complained about it) with the unlinking of the lock file. change UNLINK(ACTIVE, bufr) to UNLINK(bufr) and you should be safe ---rick
jbuck@epimass.UUCP (Joe Buck) (01/12/87)
In article <146@trwspf.UUCP> vito@trwspf.UUCP (Herb Barad) writes: >I remember reading that some people were having problems with >patch #3 that was posted for the 2.11 news. Is there any official >suggestions as to going ahead with patch #3. I think somewhere >people were mentioning that patch #3 could clear out the >active file. Is this true? If you have BSD_42 or LOCKF defined, you're safe. On other systems, the first article will wipe your active file. Here's the guilty part of the patch: + #ifdef BSD4_2 + flock(fileno(actfp), LOCK_UN); + #else /* !BSD4_2 */ + #ifdef LOCKF + lockf(fileno(actfp), F_ULOCK, 0); + #else /* !LOCKF */ + UNLINK(ACTIVE, bfr); <---- here it is! + #endif /* V7 */ + #endif /* !BSD4_2 */ That's right, folks, it unlinks your active file. Change the offending line to "UNLINK(bfr)". Note: UNLINK is a macro with no arguments; that's why the C preprocessor doesn't catch this. I'd recommend installing it with this change, since other bugs are fixed. -- - Joe Buck {hplabs,ihnp4,sun}!oliveb!epimass!jbuck HASA (A,S) Entropic Processing, Inc., Cupertino, California
badri@ur-valhalla.UUCP (Badri Lokanathan) (01/13/87)
In article <43063@beno.seismo.CSS.GOV> rick@seismo.CSS.GOV (Rick Adams) writes: > ..... >change > UNLINK(ACTIVE, bufr) >to > UNLINK(bufr) > ..... Is this posting to be treated as "official" notification (i.e patch #3a?) If a future patch comes up, shall I be expected to have modified my file or will I have to hack things again? Badri Lokanathan -- "I care about my fellow man {) ur-valhalla!badri@rochester.arpa Being taken for a ride, //\\ {ames,caip,cmcl2,columbia,cornell, I care that things start changing ///\\\ harvard,ll-xn,rutgers,seismo, But there's no one on my side."-UB40 _||_ topaz}!rochester!ur-valhalla!badri